This repo contains the code of our ICLR'21 paper, "Domain Generalization with MixStyle".
The OpenReview link is https://openreview.net/forum?id=6xHJ37MVxxp.
########## Updates ############
[06-10-2022] New paper "On-Device Domain Generalization" is out! Code, models and datasets: https://github.com/KaiyangZhou/on-device-dg.
[12-10-2021] Please note that the latest version for Dassl is v0.5.0 (the changes might affect the performance if the original images are not square). See this for more details.
[06-07-2021] Extension of our ICLR'21 paper is out: MixStyle Neural Networks for Domain Generalization and Adaptation. This work extends the conference version mainly in the following ways: 1) A simple algorithmic extension enabling MixStyle to cope with unlabeled data; 2) New evidence showing that MixStyle works exceptionally well with extremely limited labels; 3) New experiments covering semi-supervised domain generalization and unsupervised domain adaptation. Code for reproducing the new experiments is available at imcls/.
[28-06-2021] A new implementation of MixStyle is out, which merges MixStyle2 to MixStyle and switches between random and cross-domain mixing using self.mix. The new features can be found here.
[12-04-2021] A variable self._activated is added to MixStyle to better control the computational flow. To deactivate MixStyle without modifying the model code, one can do
def deactivate_mixstyle(m):
if type(m) == MixStyle:
m.set_activation_status(False)
model.apply(deactivate_mixstyle)Similarly, to activate MixStyle, one can do
def activate_mixstyle(m):
if type(m) == MixStyle:
m.set_activation_status(True)
model.apply(activate_mixstyle)Note that MixStyle has been included in Dassl.pytorch. See the code for details.
[05-03-2021] You might also be interested in our recently released survey on domain generalization at https://arxiv.org/abs/2103.02503, which summarizes the ten-year development in domain generalization, with coverage on the history, datasets, related problems, methodologies, potential directions, and so on.
##############################
A brief introduction: The key idea of MixStyle is to probablistically mix instance-level feature statistics of training samples across source domains. MixStyle improves model robustness to domain shift by implicitly synthesizing new domains at the feature level for regularizing the training of convolutional neural networks. This idea is largely inspired by neural style transfer which has shown that feature statistics are closely related to image style and therefore arbitrary image style transfer can be achieved by switching the feature statistics between a content and a style image.
MixStyle is very easy to implement. Below we show a brief implementation of it in PyTorch. The full code can be found here.
import random
import torch
import torch.nn as nn
class MixStyle(nn.Module):
"""MixStyle.
Reference:
Zhou et al. Domain Generalization with MixStyle. ICLR 2021.
"""
def __init__(self, p=0.5, alpha=0.1, eps=1e-6, mix='random'):
"""
Args:
p (float): probability of using MixStyle.
alpha (float): parameter of the Beta distribution.
eps (float): scaling parameter to avoid numerical issues.
mix (str): how to mix.
"""
super().__init__()
self.p = p
self.beta = torch.distributions.Beta(alpha, alpha)
self.eps = eps
self.alpha = alpha
self.mix = mix
self._activated = True
def __repr__(self):
return f'MixStyle(p={self.p}, alpha={self.alpha}, eps={self.eps}, mix={self.mix})'
def set_activation_status(self, status=True):
self._activated = status
def update_mix_method(self, mix='random'):
self.mix = mix
def forward(self, x):
if not self.training or not self._activated:
return x
if random.random() > self.p:
return x
B = x.size(0)
mu = x.mean(dim=[2, 3], keepdim=True)
var = x.var(dim=[2, 3], keepdim=True)
sig = (var + self.eps).sqrt()
mu, sig = mu.detach(), sig.detach()
x_normed = (x-mu) / sig
lmda = self.beta.sample((B, 1, 1, 1))
lmda = lmda.to(x.device)
if self.mix == 'random':
# random shuffle
perm = torch.randperm(B)
elif self.mix == 'crossdomain':
# split into two halves and swap the order
perm = torch.arange(B - 1, -1, -1) # inverse index
perm_b, perm_a = perm.chunk(2)
perm_b = perm_b[torch.randperm(B // 2)]
perm_a = perm_a[torch.randperm(B // 2)]
perm = torch.cat([perm_b, perm_a], 0)
else:
raise NotImplementedError
mu2, sig2 = mu[perm], sig[perm]
mu_mix = mu*lmda + mu2 * (1-lmda)
sig_mix = sig*lmda + sig2 * (1-lmda)
return x_normed*sig_mix + mu_mixHow to apply MixStyle to your CNN models? Say you are using ResNet as the CNN architecture, and want to apply MixStyle after the 1st and 2nd residual blocks, you can first instantiate the MixStyle module using
self.mixstyle = MixStyle(p=0.5, alpha=0.1)during network construction (in __init__()), and then apply MixStyle in the forward pass like
def forward(self, x):
x = self.conv1(x) # 1st convolution layer
x = self.res1(x) # 1st residual block
x = self.mixstyle(x)
x = self.res2(x) # 2nd residual block
x = self.mixstyle(x)
x = self.res3(x) # 3rd residual block
x = self.res4(x) # 4th residual block
...In our paper, we have demonstrated the effectiveness of MixStyle on three tasks: image classification, person re-identification, and reinforcement learning. The source code for reproducing all experiments can be found in mixstyle-release/imcls, mixstyle-release/reid, and mixstyle-release/rl, respectively.
Takeaways on how to apply MixStyle to your tasks:
res1 and res2 in ResNets).res1 and res2 in ResNets.For more analytical studies, please read our paper at https://openreview.net/forum?id=6xHJ37MVxxp.
Please also read the extended paper at https://arxiv.org/abs/2107.02053 for a more comprenehsive picture of MixStyle.
To cite MixStyle in your publications, please use the following bibtex entry
@inproceedings{zhou2021mixstyle,
title={Domain Generalization with MixStyle},
author={Zhou, Kaiyang and Yang, Yongxin and Qiao, Yu and Xiang, Tao},
booktitle={ICLR},
year={2021}
}
@article{zhou2021mixstylenn,
title={MixStyle Neural Networks for Domain Generalization and Adaptation},
author={Zhou, Kaiyang and Yang, Yongxin and Qiao, Yu and Xiang, Tao},
journal={arXiv:2107.02053},
year={2021}
}